Similar speaker recognition method and system using nonlinear analysis

Abstract

Disclosed herein is a similar speaker recognition method and system using nonlinear analysis. The recognition method extracts a nonlinear feature of a sound signal through nonlinear analysis of the sound signal and combines the nonlinear feature with a linear feature such as spectrum. The method transforms sound data in a time domain into status vectors in a phase domain and uses a nonlinear time series analysis method capable of representing nonlinear features of the status vectors to extract nonlinear information of a sound. The method can overcome technical limitations of conventional linear algorithms. The recognition method can be applied to sound-related application systems other than speaker recognition systems.

1 . A similar speaker recognition method, comprising the steps of:
 receiving a sound signal;   extracting a first feature from the sound signal;   extracting a second feature from the sound signal;   comparing the first feature with a prestored sound data, thereby generating a first comparing value;   comparing the second feature with the prestored sound data if the first comparing value is within a certain range, thereby generating a second comparing value; and   estimating that the sound signal and the prestored sound data are of same speaker if the second comparing value is within a threshold range,   wherein the first feature is a linear feature and the second feature is a nonlinear feature.   
   
   
       2 . The method as claimed in  claim 1 , wherein the first feature is extracted in a frequency domain and the second feature is extracted in a phase domain. 
   
   
       3 . The method as claimed in  claim 1 , wherein the first feature uses MFCC(MelFrequency Cepstrum) and the second feature uses correlation dimension. 
   
   
       4 . The method as claimed in  claim 1 , wherein a weight is applied to each of the first feature and the second feature to compare the first feature and the second feature with the prestored sound data. 
   
   
       5 . The method as claimed in  claim 1 , wherein the threshold range is a error threshold range for measuring a similarity between the second feature and the prestored sound data. 
   
   
       6 . An apparatus for similar speaker recognition, comprising:
 receiver for receiving a sound signal;   a first recognizer configured to generate a first comparing value by comparing a linear feature of the sound signal with a prestored sound data;   a second recognizer configured to generate a second comparing value by comparing a nonlinear feature of the sound signal with the prestored sound data when the first comparing value is within a certain range; and   a logic means configured to reject or allow an access by the second comparing value.   
   
   
       7 . The apparatus as claimed in  claim 6 , wherein a weight is applied to each of the first feature and the second feature to compare the first feature and the second feature with the prestored sound data.